The ray evaluator `method="ray"` does not accept `functools.partial` so the `run_function` has to be redefined as follows in the different benchmarks: ```python @property def run_function(self): def run_function_wrapper(job, **kwargs): return run_function_base(job, bb_func=ackley, **kwargs) return run_function_wrapper ``` where `run_function_base` is `from cbbo_benchmarks.utils import run_function as run_function_base`.